Novel finite element-based plate design for bridging mandibular defects: Reducing mechanical failure

Novel finite element-based plate design for bridging mandibular defects

Merema, Bram B J; Kraeima, Joep; de Visscher, Sebastiaan A H J; van Minnen, Baucke;

Spijkervet, Fred K L; Schepman, Kees-Pieter; Witjes, Max J H

Published in:

Oral diseases

DOI:

10.1111/odi.13331

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from

it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date:

2020

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Merema, B. B. J., Kraeima, J., de Visscher, S. A. H. J., van Minnen, B., Spijkervet, F. K. L., Schepman,

K-P., & Witjes, M. J. H. (2020). Novel finite element-based plate design for bridging mandibular defects:

Reducing mechanical failure. Oral diseases, 26(6), 1265-1274. https://doi.org/10.1111/odi.13331

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Oral Diseases. 2020;00:1–10. wileyonlinelibrary.com/journal/odi  |  1

1 | INTRODUCTION

Patients who require a continuity resection of the mandible due to, for example, head and neck cancer often receive a reconstruc-tion preferably including a free vascularised flap (e.g. fibula graft).

However, when a patient's general medical condition does not allow for this type of reconstructive surgery, the mandibular continuity de-fect can be bridged using solely a conventional reconstruction plate (RP). This type of RP usually needs manual bending to match the contour of the mandible. This method, however, has been reported to fail due to screw loosening or plate fracture (Gellrich et al., 2004;

Received: 14 October 2019 

|

|

O R I G I N A L A R T I C L E

Novel finite element-based plate design for bridging

mandibular defects: Reducing mechanical failure

Bram B. J. Merema  | Joep Kraeima  | Sebastiaan A. H. J. de Visscher |

Baucke van Minnen | Fred K. L. Spijkervet  | Kees-Pieter Schepman | Max J. H. Witjes

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

© 2020 The Authors. Oral Diseases published by John Wiley & Sons Ltd Department of Oral and Maxillofacial

Surgery, University Medical Center Groningen, Groningen, The Netherlands

Correspondence

Bram B. J. Merema, Department of Oral and Maxillofacial Surgery, University Medical Center Groningen, Hanzeplein 1, P.O. Box 30.001, Groningen 9700 RB, The Netherlands.

Email: b.j.merema@umcg.nl

Funding information

Province of Groningen, Groningen, The Netherlands

Abstract

Introduction: When the application of a free vascularised flap is not possible, a

seg-mental mandibular defect is often reconstructed using a conventional reconstruction plate. Mechanical failure of such reconstructions is mostly caused by plate fracture and screw pull-out. This study aims to develop a reliable, mechanically superior, yet slender patient-specific reconstruction plate that reduces failure due to these causes.

Patients and Methods: Eight patients were included in the study. Indications were as

follows: fractured reconstruction plate (2), loosened screws (1) and primary recon-struction of a mandibular continuity defect (5). Failed conventional reconrecon-structions were studied using finite element analysis (FEA). A 3D virtual surgical plan (3D-VSP) with a novel patient-specific (PS) titanium plate was developed for each patient. Postoperative CBCT scanning was performed to validate reconstruction accuracy.

Results: All PS plates were placed accurately according to the 3D-VSP. Mean 3D

screw entry point deviation was 1.54 mm (SD: 0.85, R: 0.10–3.19), and mean screw angular deviation was 5.76° (SD: 3.27, R: 1.26–16.62). FEA indicated decreased stress and screw pull-out inducing forces. No mechanical failures appeared (mean follow-up: 16 months, R: 7–29).

Conclusion: Reconstructing mandibular continuity defects with

bookshelf-recon-struction plates with FEA underpinning the design seems to reduce the risk of screw pull-out and plate fractures.

K E Y W O R D S

CAD-CAM, finite element analysis, mandibular reconstruction, mouth neoplasms, patient-specific modelling, prosthesis and implants

Katakura, Shibahara, Noma, & Yoshinari, 2004; Lopez, Dekeister, Sleiman, & Paoli, 2004; Maurer, Eckert, Kriwalsky, & Schubert, 2010; Schoning & Emshoff, 1998; Shibahara, Noma, Furuya, & Takaki, 2002). Maurer et al. (2010) describe a failure rate of 10% for screw loosening and plate fracture combined, while others report failure solely due to plate fracture in 10% of the cases (Irish et al., 1995; Schoning & Emshoff, 1998; Shibahara et al., 2002) or to screw loosening alone in 18% of the cases (Markwardt, Pfeifer, Eckelt, & Reitemeier, 2007). According to Maurer et al., (2010), all screw loos-ening occurred within the first 6 months postoperatively.

Plate fracture is seen predominantly in the regions surrounding the screws nearest to the continuity defect and can be caused by fatigue through cyclic in situ loading (Katakura et al., 2004), usually encouraged by residual stress inside the plate as a result of repetitive bending while contouring (Lindqvist et al., 2001; Martola, Lindqvist, Hanninen, & Al-Sukhun, 2007). It seems to occur mostly in segmen-tal defects that do not cross the midline and in the presence of rela-tively many remaining occlusal units (Shibahara et al., 2002). Failure of a bridging RP leads to severe discomfort and impaired oral func-tion for the patient (Lindqvist, Soderholm, Laine, & Paatsama, 1992). In most cases, an additional surgical procedure involving a secondary reconstruction is needed (Martola et al., 2007).

Numerous studies applied patient-specific reconstruction plates (PS-RP) to prevent plate fracture by changing the material or design of conventional RP (Li et al., 2013; Luo, Xu, Guo, & Rong, 2017; Narra et al., 2014; Singare, Shenggui, & Sheng, 2017) but did not look at screw pull-out and its prevention. In order to prevent failure in fur-ther developed PS-RP, it is necessary to assess current conventional reconstructions biomechanically. Therefore, this study focused on an analysis of conventional RPs with the finite elements method (FEM), to obtain insight into any weaknesses and to exclude them in further patient-specific designs.

The aim of this phase 1 study was to design and analyse PS-RPs to bridge mandibular gaps and to minimise plate fracture and screw pull-out-related failure. This was achieved through the development and clinical application of a reconstructive method using a PS-RP, based on a 3D virtual surgical planning (VSP) and FEM supported individual design.

2 | PATIENTS AND METHODS

2.1 | Patients

All the patients included in this study required a reconstruction of either a primary or a secondary mandibular continuity defect due to a fractured or loosened conventional RP. None of these patients' mandible could be reconstructed with a free vascularised bone flap due to the poor quality of the donor site vascularisation, an impaired medical condition or refusal to undergo major free vascularised bone flap surgery. The study was approved by the Medical Ethical board of our centre (no. METc-2019/301), and informed consent was ob-tained for all patients.

In the secondary cases, we started by assessing the mechanically failed conventional RPs retrospectively, following plate fracture or screw pull-out. Postoperative CT scans were used for the segmen-tation, and 3D models of the mandibular segments were obtained using Mimics 19.0 software (Materialise). In order to perform finite element analysis (FEA) on the failed primary reconstructions, we digitalised the conventional reconstruction plates, which had been bent to follow the mandible's contour. Using the postoperative CT scans of primary reconstructions, we obtained the contours the plates were bent to, and using the manufacturer's dimensions, we designed matching plates for analysis. This was necessary since the quantity of metal artefacts or scatter in the CT data would not allow for proper segmentation of the osteosynthesis material.

3D FEA (performed with Solidworks Professional 2017 soft-ware, Dassault Systèmes Solidworks Corp.) of the conventional re-constructions enabled assessment of the local stress values in the conventional reconstruction plates. The typical defect size used in our general design process was a continuity defect spanning from the mandibular angle up to the approximate midline of the mandi-ble while not crossing it. This represents a L-defect according to

F I G U R E 1   The primary and secondary reconstructions for

patient 2. Notice the under-contouring of the patient-specific secondary reconstruction plate (orange) compared to the failed primary reconstruction plate (transparent)

|

MEREMA EtAl.

Jewer's HCL classification (Jewer et al., 1989) or a class II defect according to Brown's (Brown, Barry, Ho, & Shaw, 2016) (Figure 1) and is reported to be the most prone to RP failure (Shibahara et al., 2002). In our FEA, we applied the bone material properties and the musculatory system as described by Mesnard and Ramos (Mesnard & Ramos, 2016; Mesnard et al., 2011; Ramos, Ballu, Mesnard, Talaia, & Simões, 2011; Ramos & Mesnard, 2015; Ramos, Nyashin, & Mesnard, 2017). Their models are based on in vivo muscle force measurements and musculature information derived from dissections. They were validated by comparing in silico mod-els to in vitro modmod-els of the same human mandibles. In accordance with these studies, incisal bite was simulated since this would ulti-mately load the mandible. Fixtures were applied to both condyles; thus, the lateral pterygoid muscles were not taken into account. The mandible was assumed to consist of an isotropic cancellous portion, with a cortical outer layer whereby the elastic properties are presented as Young's moduli (Young's modulus measures the stiffness of a solid material) of 400 MPa and 14,700 MPa, respec-tively (Mesnard & Ramos, 2016). Poisson's ratio, a measure of how a material constricts or expands to a tensile or compressive load, was assumed to be 0.3 for both the cortical and cancellous bone (Mesnard & Ramos, 2016). All the conventional RPs were consid-ered commercially pure grade 2 titanium, with a Young modulus of 102,000 MPa and a Poisson ratio of 0.34, while the PS-RPs and all the applied screws were assigned 113,800 MPa and 0.34 for Young's modulus and Poisson's ratio, respectively, representing ti-tanium grade 5 alloy (Boyer, Welsch, & Collings, 1994; Holt & Ho, 1996). We used this titanium alloy because of its higher resistance against fatigue, or endurance limit.

2.2 | Design and production

After analysing the conventional RP, an alternative PS reconstruc-tion method was developed consisting of a 3D virtual surgical plan (VSP) that combines both CT data (i.e. bone segmentation and nerve canal delineation) and fused with MRI-based tumour delineation for

osteotomy placement (Kraeima, Dorgelo, et al., 2018). The aim was to overcome screw pull-out and high stresses which could lead to plate fracture.

Contrary to the majority of PS-RP suggested in the literature, which typically consist of a strip-like plate following the buccal contour of the mandible, we focused on incorporating the oste-otomy sites of the mandibular segments for stable fixation of the plate. Bookshelf-like flanges situated against the osteotomy planes of the mandibular segments were added to a bridging section. To mitigate the chance of dehiscence of the plate due to contraction of the covering soft tissues, the bridging section was under-con-toured with respect to the lateral and caudal boundaries of the preresected mandible (Figure 1). Fixation of the plate was ob-tained through bi-cortical screw placement and, whenever possi-ble, mono-cortically in the flanges supporting the osteotomy sites. The design was carried out in 3-Matic Medical 11.0 (Materialise). Once finished, the STL file of the plate was exported and con-verted into a non-uniform rational basis spline (NURBS) object using Geomagic software (3D Systems). Subsequently, screw threads, compatible with 2.3 locking screws (KLS Martin), were in-serted into the NURBS file using the Solidworks Professional 2017 software. Subsequently, the finalised CAD files were sent to the manufacturer (Witec Fijnmechanische Techniek BV) to mill the PS plates from medical grade 5 titanium alloy.

In order to accurately translate the VSP to the operating the-atre, surgical drill and cutting guides were designed in house and 3D printed by Oceanz (Oceanz BV) from medical grade polyamide powder (Figure 2). Cylinders in the guide indicate the position and direction of the planned screws and function as a pilot drill support when an additional metallic drill sleeve is inserted. Comparable guide designs were applied in previous studies and proved to be an ac-curate translator of surgical plans (Kraeima, Dorgelo, et al., 2018; Kraeima, Merema, Witjes, & Spijkervet, 2018; Schepers et al., 2015, 2016; Vosselman, Merema, Schepman, & Raghoebar, 2018). In addi-tion, centring pins were designed to function as intraoperative plate positioning devices and inserted into the drilled screws pilot holes (Figure 2).

F I G U R E 2   This figure shows an

example of the type of surgical guides (top right) and temporary centring pins (lower right) that were used in this study

T A B LE 1  Pa tien t o ver vi ew Pa tie nt A ge Sex D ia gn osi s In dic at io n PM f la p Fo llo w -u p (mo nt hs ) N um be r o f sc rew s Sc re w e nt ry po in t d ev . ( m m ) Sc re w a ng ul ar de v. ( °) 1 49 ♂ M et as ta si s me la no m a m an dib le B ro ke n 2 .7 O SM p late No 29 13 2. 20 (± 0. 51 ) 5. 41 (± 3. 32 ) 2 85 ♂ T4 N 0 S C C B ro ke n 2 .7 O SM p late Ye s 25 9 1. 49 (± 0. 42 ) 4. 81 (± 2. 56 ) 3 58 ♀ T4 N 0 S C C Pr im ar y re co n. Ye s 21 9 1. 31 (± 0. 30 ) 5. 99 (± 3. 52 ) 4 72 ♀ T4 N 0 S C C Pr im ar y re co n. Ye s 18 8 0. 35 (± 0. 22 ) 4. 26 (± 2. 29 ) 5 85 ♂ T4N x S C C Pr im ar y re co n. Ye s NA 8 NA NA 6 70 ♂ A m elo bl as to m a m an dib le Lo ose ne d sc re w s 2 .7 O SM p late s No 8 8 2. 04 (± 0. 75 ) 7. 49 (± 4. 27 ) 7 80 ♀ T4 N 0 S C C m an dib le Pr im ar y re co n. Ye s 7 7 0. 87 (± 0. 59 ) 4. 94 (± 3. 46 ) 8 78 ♂ M et as ta si s T2 N 2 S C C m an dib le Pr im ar y re co n. Ye s 7 10 1. 83 (± 0. 96 ) 6. 29 (± 3. 20 ) M ea n 72 16 ( R: 7– 29 ) 9 ( R: 7 –1 3) 1. 54 (± 0. 85 ) 5. 76 (± 3. 27 ) A bb rev iat io ns : ♀ , f em al e; ♂ , m al e; O SM , o st eo sy nt he si s m at er ia l; P M , p ec to ra lis m aj or ; S C C , s qu am ou s c el l c ar ci no m a.

|

MEREMA EtAl.

2.3 | Measurements

All the patients underwent a postoperative CBCT scan (120 kV/5 mA with a field of view of 130–230 mm and 0.2–0.4 mm voxel size) to assess the accuracy of implant placement by means of screw entry point deviation and angular screw deviation. Manual alignment of the planned 3D objects of the mandibular segments and screw cyl-inders with the postoperative CBCT was performed with the Mimics Medical 19.0 software. Two observers executed the alignment inde-pendently (BJM and JK). The in situ plate was segmented in order to assess angular screw deviation and subsequently matched to the plate's design file, while the manually aligned screw cylinders were moved along. The Geomagic Studio 2012 (3D Systems) soft-ware was used for the matching through a best-fit surface alignment procedure.

Screw entry point deviation was measured between the entry points in the virtual planning, and the cylinders were matched to the postoperative CBCT by means of Euclidean dis-tance (3D) measurements in the 3-Matic Medical 11.0 software. Prior to these measurements, all the manually aligned mandibu-lar segments and corresponding cylinders were matched to the virtual planning using the global alignment function in 3-Matic Medical 11.0.

Data analysis was performed using MedCalc for Windows, version 19.0.5 (MedCalc Software). The inter-observer variability was supported by the calculation of the interclass correlation coefficient (ICC) for every screw entry point, placed by both observers. A value of <.40 is reported as poor, .40–.59 fair, .60–.74 good and .75–1.00 as excellent (Cicchetti, 1994).

3 | RESULTS

3.1 | Patients

A total of eight patients who had either already undergone recon-struction of a continuity resection of the mandible or were sched-uled to undergo one presented to our centre (n = 8). This group required either primary treatment of tumours (n = 5) or replacement of a mechanically failed conventional reconstruction plate (n = 3). The latter consisted of two patients with a broken conventional 2.7 RP and one patient with pulled out screws, causing loosening of the conventional 2.7 RP and the locking screws. Table 1 shows the detailed overview of these patients. Patients 7 and 8 received postoperative radiotherapy (66 Gy), starting within six weeks after reconstruction with the PS-RP.

We created comparative FEA of our PS reconstructions for the three patients with failed hand-bent reconstruction plates. Mandibular segments, loading situations and boundary conditions remained unchanged. The comparative FEA considered the resultant forces on the bone–screw interface, as well as the von Mises or re-sultant stress occurring in the reconstruction plates. This von Mises

stress was used to predict whether or not materials will yield under loading.

3.2 | Plate fracture

The FEA results showed that the maximum von Mises stresses in all the analysed conventional RPs exceeded their yield strength (YS), by 42% up to 153%, indicating plastic deformation would occur on ap-plying the load case. The plates would therefore not return to their original shape after loading. Furthermore, the stress in these plates exceeded the material's ultimate strength (US) value by 13% up to 100%, indicating a high risk of plate fracture. The application of our PS reconstruction to the latter patient resulted in a decrease in the YS and US percentages, going from 253% to 59% and 200% to 55%, respectively (Figure 3). In patient number two, who suffered from a fractured conventional RP, we saw a decrease in YS and US per-centages, going from 188% to 67% and 149% to 62%, respectively (Figure 4), thereby staying well within acceptable boundaries.

3.3 | Screw pull-out

The results of our FEA of the primary conventional reconstruction that failed due to screw pull-out showed that high resultant forces were acting on these screw surrounding areas. In this case, the RP was fixed at the ventral side of the defect with three screws. Screw one, two and three, counting away from the defect, which had loos-ened over time, were loaded with 438 N, 416 N and 112 N, respec-tively. The comparative FEA of our bookshelf-plate design showed that adding a screw in the ventral bookshelf flange (205 N) would lower the resultant screw forces to 184 N, 144 N and 90 N, respec-tively, which represents a minimisation of 58%, 66% and 20%. The axial pull-out force components, the forces along the longitudinal screw directions, did not exceed 180 N for the conventional recon-structions or 90 N for the PS reconrecon-structions.

3.4 | Surgical procedure

All the PS bookshelf-reconstruction plates were inserted in accord-ance with the 3D-VSP. The surgical procedures were uneventful. During surgery, prior to the drilling of screw pilot holes, the man-dibular bone was denuded and the guides were positioned and fixed using 1.5-mm mini screws (KLS Martin). Subsequently, the osteoto-mies were performed. Thereafter, the plate was inserted and fitted to the mandibular segments using several centring pins. Once prop-erly aligned, these pins were replaced one by one by 2.3-mm lock-ing screws with a length in agreement with the surgical plannlock-ing. Primary closure was performed according to plan in two patients. In the remaining six patients, a pectoralis major flap was used to recon-struct the soft tissue defect and to cover the plate.

3.5 | Postoperative

Recovery was uneventful from a mechanical point of view for seven patients with a mean of 11.4 days of hospitalisation (SD: 9.3, R: 3–29). One patient, however, patient 5 in Table 1, deceased in the fourth week postoperatively due to complications related to a PRG probe and therefore could not be followed up. All pa-tients underwent a CBCT scan 6–27 days (mean 12 days) post-operatively. The mean follow-up period of the 7 patients alive is 16 months (R: 7–29) and was uneventful with regard to plate fail-ure or screw pull-out.

3.6 | Measurements

Screw entry point deviation (3D) resulted in a mean value of 1.54 mm (SD 0.85, R: 0.10–3.19) for a total of 64 screws. The mean angular screw deviation was 5.76 degrees (SD: 3.27, R: 1.26–16.62). The 95% confidence interval of the inter-observer variability for our measure-ments was 0.15–0.26 mm with a P-value of .05. The interclass corre-lation coefficient (two-way mixed) was 0.97, indicating an excellent match of measurements by both observers.

3.7 | Dehiscence

The ventral bookshelf-like flange of our PS-RPs in the first and third op-erated patients became partially dehiscent intra-oral approximately

14 and 4 months postoperatively, respectively. Patient 1 lost 16 kg of bodyweight over a short period of time prior to the intra-oral de-hiscence, which could have played a role in this development. One of these two patients had received a pectoralis major flap, while the other patient underwent primary closure. It was assessed that the design of the flange was too high. The dehiscent cranial part of this flange was surgically removed with some margin (23 and 19 months postoperatively, respectively), and the surrounding soft tissue could be closed (Figure 5). By comparing the 6 days postoperative panorex image of patient 1 to its 18-month follow-up, a gradual resorption of the left mandibular angle up to the caudal contour of the plate was observed. There have been no further complications to date.

4 | DISCUSSION

In this pilot study, we present a unique patient-specific bookshelf-reconstruction plate for accurate bridging of mandibular con-tinuity defects. The novel design was based on the FEA of the conventional reconstruction plate failure, with regard to plate fractures and screw loosening. Application of bookshelf-like flanges, with a screw fixation in the osteotomy sites, resulted in substantial reduction of resultant screw pull-out inducing forces and, in combination with a change in material, lowered plate stress to within satisfactory levels. Therefore, the chances of the widely reported mechanical problems seen with conventional RPs, which are responsible for 5%–10% of reconstruction failures, were re-duced (Gellrich et al., 2004; Katakura et al., 2004; Lopez et al.,

F I G U R E 3   Finite element analysis

shows the maximum occurring von Mises stress in both the failed conventional (primary) reconstruction plate (top left) and the PS reconstruction plate (lower left) in patient one and illustrates the resemblance between the highest stress region of the conventional reconstruction plate in silico (note the overloaded pink region in the top left image) and the actual location of the in situ plate fracture (panorex)

|

MEREMA EtAl.

2004; Maurer et al., 2010; Schoning & Emshoff, 1998; Shibahara et al., 2002).

Our comparative FEA focused on resultant screw forces rather than the force components in axial screw direction. Most studies found in the literature that look into screw pull-out describe in vitro results of axial single screw pull-out. This observation of axial pull-out could occur with compression screws but not with locking screws, since these have the tendency to be pulled out en bloc with the RP rather than axially, due to their semi-rigid connection with the RP (Cronier et al., 2010). We do know, however, that the axial force components regarding the screws in our study did not exceed 180 N for the conventional reconstructions and 90 N for the PS reconstructions.

The addition of our bookshelf-flange concept is only of value, in terms of mechanical stability and reduction of the risk of fail-ure, when applied accurately to the planned position. First, a per-operative visual inspection after guided placement confirmed that no gap remained between the flange and the mandibu-lar bone at the osteotomy sites. Second, postoperative analysis of CBCT scans showed a high accuracy of placement, with 3D deviations comparable to our prior studies and of others, using

surgical guides (Kraeima, Dorgelo, et al., 2018; Kraeima, Merema, et al., 2018; Pietruski et al., 2019; Schepers et al., 2016; van Baar, Forouzanfar, Liberton, Winters, & Leusink, 2018; Vosselman et al., 2018). Additionally, careful inspection of the postoperative CBCT scans confirmed contact between the osteotomy sites and book-shelf-like flanges of the RPs as well as the remainder bone–plate interface, which indicates proper positioning with a small potential shift in the osteotomy plane. The measured screw entry point de-viation of 1.58 mm (SD 0.82) and angular screw dede-viation of 5.77 degrees (SD 3.33) indicate this method could be applied as a reli-able one-phase procedure for resection and direct reconstruction of a tumour in the mandible. These results represent the accu-mulation of errors in all visualisation and segmentation steps as well as geometrical errors of the guides and plate and the actual surgical procedure.

Over the last two decades, a rapidly increasing number of PS-RPs have been presented in the literature with an equal increase in vary-ing finite element analysis models. Applyvary-ing FEA is of great impor-tance in the design process, since PS-RPs can still fail mechanically when designed using incorrect assumptions (Li et al., 2014; Luo et al., 2017). Only a very small selection of these FEA models has actually

F I G U R E 4   Finite element analysis

von Mises stress results for the failed conventional reconstruction (top left) and patient-specific solution (lower left) for patient 2. The Bridging part height was reduced compared to the first application of our patient-specific plate, as shown in Figure 2, and remained at this height throughout this study's series

been validated through in vitro and in vivo experiments. Engineers should be always careful when setting up a FEA model, especially complex anatomical models, and all the necessary assumptions that come with it. Also, PS-RP's should be designed to withstand repeti-tive loading and the material's fatigue properties for FEA should be chosen accordingly, like we did in this study. Often, only the ultimate or yield properties of a material are taken into account, while the fa-tigue properties are lower, which could lead to early material failure. We decided to use the most extensively validated model we could find in the literature. However, even this model has its limitations and assumptions. We found that most of the PS-RPs designs in the literature rely on a strip-like plate which is positioned and fixated

at the buccal contour of the mandible (Gutwald, Jaeger, & Lambers, 2017; Li et al., 2014; Luo et al., 2017; Mazzoni et al., 2013; Narra et al., 2014; Wu, Lin, Liu, & Lin, 2017). We decided to make use of the osteotomy site as well and trap the mandibular segment in between the plate design, which is more stable biomechanically, according to our FEA results.

We expect the dehiscence of the cranial section of the ventral flanges, that occurred in two patients, to be caused by the height of the flanges in combination with contraction of the covering soft tissue. Minimising the bulkiness and height of these flanges might exclude this occurrence. A gradual remodelling of the mandibular angle was observed in one patient past the implant border. The re-construction plate could have shielded this particular region of the mandible mechanically, causing bone remodelling to occur. Stress-shielding, also seen in conventional reconstruction plate recon-structions, might explain this resorption. Future application of the topology optimisation technique (Iqbal et al., 2019) could play a key role in minimising the occurrence of, or potentially totally exclude, both stress-shielding and dehiscence. This engineering technique removes unloaded or unnecessary material and is applied in the FEA phase of the design. It could be used to create geometrically minimalistic designs, while approaching displacement or stress lim-its, and could prevent a RP from being too stiff, which can cause stress-shielding, and plates becoming bulky. This study describes a first step in the optimisation of patient-specific plate design. By allowing for freeform organic structures through topology optimis-ation, we expect we can lift our patient-specific reconstructions to a higher level of patient specificity in the near future. Also, based on the results of this phase-one study we aim to start a multi-centre phase-two study in which we can further validate the effect of our reconstruction method in more patients and over a longer follow-up period.

5 | CONCLUSION

Using the finite element method, we retrospectively analysed me-chanically failed conventional reconstruction plates and developed an alternative reconstruction plate for the mandible that reduces the chance of screw pull-out and plate fracture. During this phase-one study, we successfully reconstructed mandibular continuity defects using our bookshelf-reconstruction plate concept in eight patients and no mechanical failures have occurred in the study cohort. This novel design of reconstructive plates seems to reduce the risk of screw pull-out and plate fractures.

ACKNOWLEDGEMENTS

The materials used in this study were unrestrictedly supported by an Innovative Action programme Groningen-4 (IAG4) grant from the Province of Groningen, The Netherlands.

CONFLIC T OF INTEREST

There is no conflict of interest to report.

F I G U R E 5   Dehiscence of the cranial section of the ventral

bookshelf-flange in patient 3. The top and middle images are intra-oral and panorex images, respectively, before surgical bookshelf-flange modification. The lower panorex image shows the postoperative situation

|

MEREMA EtAl.

AUTHOR CONTRIBUTIONS

All authors were involved in the study design and critical reading of the manuscript. Merema and Kraeima analysed the data and Merema, Kraeima and Witjes dr.

ORCID

Bram B. J. Merema https://orcid.org/0000-0003-0994-3104 Joep Kraeima https://orcid.org/0000-0001-8527-960X Fred K. L. Spijkervet https://orcid.org/0000-0003-0200-2344

REFERENCES

Boyer, R., Welsch, G., & Collings, E. W. (1994). Materials properties hand-book: Titanium alloys. Novelty, OH: ASM International.

Brown, J. S., Barry, C., Ho, M., & Shaw, R. (2016). A new classification for mandibular defects after oncological resection. Lancet Oncology, 17, 23. https://doi.org/10.1016/S1470 -2045(15)00310 -1

Cicchetti, D. (1994). Guidelines, criteria, and rules of thumb for eval-uating normed and standardized assessment instrument in psy-chology. Psychological Assessment, 6, 284–290. https://doi. org/10.1037/1040-3590.6.4.284

Cronier, P., Pietu, G., Dujardin, C., Bigorre, N., Ducellier, F., & Gerard, R. (2010). The concept of locking plates. Orthopaedics & Traumatology, Surgery & Research, 96, S17–S36. https://doi.org/10.1016/j. otsr.2010.03.008

Gellrich, N. C., Suarez-Cunqueiro, M. M., Otero-Cepeda, X. L., Schon, R., Schmelzeisen, R., & Gutwald, R. (2004). Comparative study of locking plates in mandibular reconstruction after ablative tumor surgery: THORP versus UniLOCK system. Journal of Oral and Maxillofacial Surgery, 62(2), 186–193. https://doi.org/10.1016/j. joms.2003.04.012

Gutwald, R., Jaeger, R., & Lambers, F. M. (2017). Customized man-dibular reconstruction plates improve mechanical performance in a mandibular reconstruction model. Computer Methods in Biomechanics and Biomedical Engineering, 20(4), 426–435. https://doi. org/10.1080/10255 842.2016.1240788

Holt, J. M., & Ho, C. Y. (1996). Structural alloys handbook. West Lafayette, IN: CINDAS/PURDUE University.

Iqbal, T., Wang, L., Li, D., Dong, E., Fan, H., Fu, J., & Hu, C. (2019). A general multi-objective topology optimization methodology de-veloped for customized design of pelvic prostheses. Medical Engineering & Physics, 69, 8–16. https://doi.org/10.1016/j.meden gphy.2019.06.008

Irish, J. C., Gullane, P. J., Gilbert, R. W., Brown, D. H., Birt, B. D., & Boyd, J. B. (1995). Primary mandibular reconstruction with the titanium hol-low screw reconstruction plate: Evaluation of 51 cases. Plastic and Reconstructive Surgery, 96(1), 93–99.

Jewer, D. D., Boyd, J. B., Manktelow, R. T., Zuker, R. M., Rosen, I. B., Gullane, P. J., … Freeman, J. E. (1989). Orofacial and mandibular re-construction with the iliac crest free flap: A review of 60 cases and a new method of classification. Plastic and Reconstructive Surgery, 84(3), 391–395.

Katakura, A., Shibahara, T., Noma, H., & Yoshinari, M. (2004). Material analysis of AO plate fracture cases. Journal of Oral and Maxillofacial Surgery, 62(3), 348–352. https://doi.org/10.1016/j.joms.2003.05.009 Kraeima, J., Dorgelo, B., Gulbitti, H. A., Steenbakkers, R., Schepman, K. P., Roodenburg, J., … Witjes, M. (2018). Multi-modality 3D mandib-ular resection planning in head and neck cancer using CT and MRI data fusion: A clinical series. Oral Oncology, 81, 22–28. https://doi. org/10.1016/j.oralo ncolo gy.2018.03.013

Kraeima, J., Merema, B. J., Witjes, M. J. H., & Spijkervet, F. K. L. (2018). Development of a patient-specific temporomandibular joint prosthe-sis according to the groningen principle through a cadaver test series.

Journal of Cranio-Maxillo-Facial Surgery, 46(5), 779–784. https://doi. org/10.1016/j.jcms.2018.03.004

Li, P., Shen, L., Li, J., Liang, R., Tian, W., & Tang, W. (2014). Optimal de-sign of an individual endoprosthesis for the reconstruction of ex-tensive mandibular defects with finite element analysis. Journal of Cranio-Maxillo-Facial Surgery, 42(1), 73–78. https://doi.org/10.1016/j. jcms.2013.02.005

Li, P., Tang, Y., Li, J., Shen, L., Tian, W., & Tang, W. (2013). Establishment of sequential software processing for a biomechanical model of mandibular reconstruction with custom-made plate. Computer Methods and Programs in Biomedicine, 111(3), 642–649. https://doi. org/10.1016/j.cmpb.2013.05.024

Lindqvist, C., Soderholm, A. L., Laine, P., & Paatsama, J. (1992). Rigid re-construction plates for immediate rere-construction following mandib-ular resection for malignant tumors. Journal of Oral and Maxillofacial Surgery, 50(11), 1158–1163. 0278-2391(92)90146 -Q

Lindqvist, C., Soderholm, A. L., Salo, A., Subasinghe, J., Ylijoki, S., Skutnabb, K., & Hallikainen, D. (2001). A comparative study on four screw-plate locking systems in sheep: A clinical and radiological study. International Journal of Oral and Maxillofacial Surgery, 30(2), 160–166. https://doi.org/10.1054/ijom.2000.0037

Lopez, R., Dekeister, C., Sleiman, Z., & Paoli, J. R. (2004). Mandibular reconstruction using the titanium functionally dynamic bridging plate system: A retrospective study of 34 cases. Journal of Oral and Maxillofacial Surgery, 62(4), 421–426. https://doi.org/10.1016/j. joms.2003.06.011

Luo, D., Xu, X., Guo, C., & Rong, Q. (2017). Fracture prediction for a customized mandibular reconstruction plate with finite element method. In Paper presented at the Advanced Computational Methods in Life System Modeling and Simulation (pp. 86–94).

Markwardt, J., Pfeifer, G., Eckelt, U., & Reitemeier, B. (2007). Analysis of complications after reconstruction of bone defects involving com-plete mandibular resection using finite element modelling. Onkologie, 30(3), 121–126. 98848

Martola, M., Lindqvist, C., Hanninen, H., & Al-Sukhun, J. (2007). Fracture of titanium plates used for mandibular reconstruction following ab-lative tumor surgery. Journal of Biomedical Materials Research. Part B, Applied Biomaterials, 80(2), 345–352. https://doi.org/10.1002/ jbm.b.30603

Maurer, P., Eckert, A. W., Kriwalsky, M. S., & Schubert, J. (2010). Scope and limitations of methods of mandibular reconstruction: A long-term follow-up. British Journal of Oral & Maxillofacial Surgery, 48(2), 100–104. https://doi.org/10.1016/j.bjoms.2009.07.005

Mazzoni, S., Marchetti, C., Sgarzani, R., Cipriani, R., Scotti, R., & Ciocca, L. (2013). Prosthetically guided maxillofacial surgery: Evaluation of the accuracy of a surgical guide and custom-made bone plate in oncology patients after mandibular reconstruction. Plastic and Reconstructive Surgery, 131(6), 1376–1385. https://doi.org/10.1097/PRS.0b013 e3182 8bd6b0

Mesnard, M., & Ramos, A. (2016). Experimental and numerical predic-tions of biomet((R)) alloplastic implant in a cadaveric mandibular ramus. Journal of Cranio-Maxillo-Facial Surgery, 44(5), 608–615. https://doi.org/10.1016/j.jcms.2016.02.004

Mesnard, M., Ramos, A., Ballu, A., Morlier, J., Cid, M., & Simoes, J. A. (2011). Biomechanical analysis comparing natural and alloplastic temporomandibular joint replacement using a finite element model. Journal of Oral and Maxillofacial Surgery, 69(4), 1008–1017. https:// doi.org/10.1016/j.joms.2010.02.019

Narra, N., Valasek, J., Hannula, M., Marcian, P., Sandor, G. K., Hyttinen, J., & Wolff, J. (2014). Finite element analysis of customized recon-struction plates for mandibular continuity defect therapy. Journal of Biomechanics, 47(1), 264–268. https://doi.org/10.1016/j.jbiom ech.2013.11.016

Pietruski, P., Majak, M., Swiatek-Najwer, E., Zuk, M., Popek, M., Mazurek, M., Jaworowski, J. (2019). Supporting mandibular resection with

intraoperative navigation utilizing augmented reality technology – A proof of concept study. Journal of Cranio-Maxillo-Facial Surgery, 47(6), 854–859. https://doi.org/10.1016/j.jcms.2019.03.004

Ramos, A., Ballu, A., Mesnard, M., Talaia, P., & Simões, J. (2011). Numerical and experimental models of the mandible. Experimental Mechanics, 51(7), 1053–1059. https://doi.org/10.1007/s1134 0-010-9403-x Ramos, A. M., & Mesnard, M. (2015). The stock alloplastic

temporoman-dibular joint implant can influence the behavior of the opposite na-tive joint: A numerical study. Journal of Cranio-Maxillofacial Surgery, 43(8), 1384–1391. https://doi.org/10.1016/j.jcms.2015.06.042 Ramos, A., Nyashin, Y., & Mesnard, M. (2017). Influences of

geometri-cal and mechanigeometri-cal properties of bone tissues in mandible behaviour - experimental and numerical predictions. Computer Methods in Biomechanics and Biomedical Engineering, 20(9), 1004. https://doi. org/10.1080/10255 842.2017.1322072

Schepers, R. H., Kraeima, J., Vissink, A., Lahoda, L. U., Roodenburg, J. L. N., Reintsema, H., … Witjes, M. J. (2016). Accuracy of secondary max-illofacial reconstruction with prefabricated fibula grafts using 3D planning and guided reconstruction. Journal of Cranio-Maxillo-Facial Surgery, 44(4), 392–399. https://doi.org/10.1016/j.jcms.2015.12.008 Schepers, R. H., Raghoebar, G. M., Vissink, A., Stenekes, M. W., Kraeima,

J., Roodenburg, J. L., … Witjes, M. J. (2015). Accuracy of fibula re-construction using patient-specific CAD/CAM rere-construction plates and dental implants: A new modality for functional reconstruction of mandibular defects. Journal of Cranio-Maxillo-Facial Surgery, 43(5), 649–657. https://doi.org/10.1016/j.jcms.2015.03.015

Schoning, H., & Emshoff, R. (1998). Primary temporary AO plate re-construction of the mandible. Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontics, 86(6), 667–672. https:// doi.org/10.1016/S1079 -2104(98)90201 -3

Shibahara, T., Noma, H., Furuya, Y., & Takaki, R. (2002). Fracture of mandibular reconstruction plates used after tumor resection. Journal of Oral and Maxillofacial Surgery, 60(2), 182–185. https://doi. org/10.1053/joms.2002.29817

Singare, S., Shenggui, C., & Sheng, L. (2017). The use of 3D printing technology in human defect reconstruction-a review of cases study. Medical Research and Innovations, 1(2), 1–4. https://doi.org/10.15761 /MRI.1000109

van Baar, G. J. C., Forouzanfar, T., Liberton, N. P. T. J., Winters, H. A. H., & Leusink, F. K. J. (2018). Accuracy of computer-assisted surgery in mandibular reconstruction: A systematic review. Oral Oncology, 84, 52–60. https://doi.org/10.1016/j.oralo ncolo gy.2018.07.004 Vosselman, N., Merema, B. J., Schepman, K. P., & Raghoebar, G. M.

(2018). Patient-specific sub-periosteal zygoma implant for prosthetic rehabilitation of large maxillary defects after oncological resection. International Journal of Oral and Maxillofacial Surgery, 48, 115–117. Wu, C. H., Lin, Y. S., Liu, Y. S., & Lin, C. L. (2017). Biomechanical evaluation

of a novel hybrid reconstruction plate for mandible segmental de-fects: A finite element analysis and fatigue testing. Journal of Cranio-Maxillo-Facial Surgery, 45(10), 1671–1680. https://doi.org/10.1016/j. jcms.2017.07.010

SUPPORTING INFORMATION

Additional supporting information may be found online in the Supporting Information section.

How to cite this article: Merema BBJ, Kraeima J, de Visscher

SAHJ, et al. Novel finite element-based plate design for bridging mandibular defects: Reducing mechanical failure. Oral Dis. 2020;00:1–10. https://doi.org/10.1111/odi.13331